← knowledge.oriz.in

Testing services

index servicestestingvitestplaywrightstorybookchromaticmswmockoonciindex

Testing services

The family runs a three-layer testing stack on every PR. Each layer answers a different question; together they form a defensive net comparable to the three-tool a11y stack.

API mocks split by surface (locked 2026-06-20 — api-mocks-msw-plus-mockoon):

Service Status One-line role
vitest.md active Unit + integration runner — Vite-native, free OSS
playwright.md active E2E + cross-browser automation — also runs the a11y axe-core suite
storybook.md active Component sandbox — feeds Chromatic snapshots, also doubles as visual docs
chromatic.md active Visual regression diff against Storybook — 5K snapshots/mo free
msw.md active In-process API mocks — Vitest + Storybook + dev (Service Worker / Node interceptor)
mockoon.md active Out-of-process API mocks — E2E + manual dev against third-party APIs (Razorpay sandbox, data APIs offline)

Why all three (four)?

PR fails on any failure in any layer. The policy is locked in decisions/architecture/testing-three-layer.md.

Why two API-mock tools?

Different surfaces, different process models — same posture as the AI split:

Both free, both OSS. Locked in api-mocks-msw-plus-mockoon.

Cross-refs